US Leans Toward Targeted Bans on Chinese Open-Weight Models, Not a Blanket Prohibition

The Trump administration is leaning toward targeted bans on individual Chinese open-weight AI models rather than a blanket prohibition, The Decoder reported in July 2026, citing national security concerns.

What exactly does the US plan to do?

The US is weighing a ban on specific Chinese open-weight models, not on all open-weight systems at once. According to The Decoder, the Trump administration sees a selective approach as less disruptive to the market and less vulnerable to accusations of blanket protectionism than a sweeping barrier.

• The restrictions would be targeted, applied to individual models rather than a general ban on the entire open-weight class
• The stated reason is national security
• OpenAI and Google DeepMind signed an open letter against regulating open-weight models
• OpenAI and Anthropic, according to the source, are privately lobbying for the same restrictions
• Status: reported as a plan (reportedly), not as a decision that has been made

Why does OpenAI's position look contradictory?

OpenAI is positioning itself both against and in favor of regulating open-weight models. Publicly, the company signed an open letter with Google DeepMind opposing such restrictions, a response to public pressure. Privately, however, The Decoder reports that OpenAI and Anthropic continue to lobby for the very same bans.

According to

The Decoder, the Trump administration is leaning toward banning specific Chinese models rather than all open-weight systems at once, citing security risks.

The gap between the public and behind-the-scenes positions reflects a dual motivation. On one side are the security risks of foreign models. On the other are the direct business interests of American labs, which stand to gain from restricting strong Chinese competitors such as DeepSeek and Qwen.

What is behind the decision?

Two factors drive the choice of a targeted approach: security and competition. Open-weight models can be downloaded and run locally without the developer's control, so regulators worry that American companies and government agencies will adopt Chinese open-source models en masse.

At the same time, restricting Chinese models plays into the hands of OpenAI, Anthropic and Google by easing pressure from cheap, powerful open competitors. It is this overlap between security arguments and commercial benefit that makes the July 2026 lobbying especially conspicuous.

What does this mean?

Washington is feeling its way toward a middle path between the freedom of the open-weight ecosystem and a hard ban: hit specific Chinese models rather than the whole class. For the industry, this is a signal that regulating open models is no longer a theoretical debate and is becoming an instrument of technological rivalry between the US and China.

Frequently asked questions

Does the US plan to ban Chinese AI models entirely?

No. According to The Decoder, the Trump administration is leaning toward targeted bans on individual Chinese open-weight models, not a blanket ban on the entire open-weight class.

Why does OpenAI's position look contradictory?

OpenAI, together with Google DeepMind, publicly signed an open letter against regulating open-weight models but, according to the source, is privately continuing to lobby for the same restrictions alongside Anthropic.